Asklemmy

Yes and no. Some people think that the zipper merge would improve traffic flow, and eliminate backup at a merge point. I say that it's not failure to zipper-merge that causes backups, but it's backups the provide the conditions in which people fail to zipper-merge.

We can divide traffic at a 2-lane-to-1-lane merge into three conditions: Traffic volume that 1 lane can handle; traffic volume that's too much for 1 lane, but 2 lanes can handle; and traffic volume that's too much for 2 lanes. Obviously, too much traffic for 2 lanes, in 2 lanes, is already a traffic jam, and a zipper merge cannot fix it. Traffic volume appropriate for 1 lane, but that's spread out over 2 lanes can easily coalesce into 1 lane. The zipper merge is good and appropriate in this scenario, and this is where I've seen it work the way some people think it should. But that's not zipper-merge magic. Merging early works, too, because a volume of traffic that fits in 1 lane, well, fits in 1 lane.

It's the scenario in which there's traffic that fits in 2 lanes has to merge into 1 lane that's the problem. Just like how you can't fit a water buffalo in your carry-on bag, no matter what clever packing technique you employ, there are just too many cars for 1 lane, and there's going to be a backup. It's only after there's a backup that the conflict between the merge-early and zipper-merge people arises, so picking one or the other can't change the basic fact of the traffic backup. At best, the zipper merge reduces the spatial length of the queue of cars waiting to merge, but does not improve the throughput of vehicles at the merge point.